Available studies suggest that indoor air pollution from cooking and space heating causes substantial ill health in developing countries where a majority of households rely on unprocessed biomass fuels such as wood, crop residues, and dung cakes, but research on this subject is limited and many uncertainties remain. The objective of the proposed research is to assess the effects of exposure to cooking smoke on a number of health outcomes including tuberculosis, asthma, acute respiratory infections, anemia, and adverse pregnancy outcomes. The primary data source for the project is India's 1998-99 National Family Health Survey (NFHS-2), comprising more than 90,000 households throughout the country. This survey, which two members of the proposed research team helped design, contains detailed information on cooking fuels, fuel mix, tobacco smoking, and a number of relevant health outcomes, as well as information on many individual and household characteristics that may potentially confound the relationships between cooking smoke, tobacco smoke, and health. The research will also utilize data from India's 1992-93 National Family Health Survey (NFHS-1), which also covered about 90,000 households throughout India; but the use of NFHS-1 data will be limited because it did not collect data on asthma, anemia, and some other key variables including tobacco smoking. Data from these surveys are unique in that they include questions on both cooking fuel type and a number of health outcomes, as well as data on tobacco smoking, from a large national sample, making it possible to explore the relationships between cooking smoke, tobacco smoke, and these health outcomes. The principal method of analysis will be logistic regression, with presence or absence of a particular health condition (e.g., active tuberculosis) as the response variable. The principal predictor variable will be exposure to cooking smoke, measured indirectly by type of cooking fuel used by the household (biomass fuels, cleaner fuels, mix of biomass and cleaner fuels). The regressions will include a number of demographic, socioeconomic, and environmental control variables, including both active and passive exposure to tobacco smoke.